Like many neural circuits of the brain, the retina is composed of a network of cells with greatly varying anatomical and physiological properties. The most diverse types of cells, both in the retina and cortex, are inhibitory interneurons. Retina amacrine cells comprise over thirty types and influence the responses of ganglion cells, the output cells of the retina. Although the anatomy and physiology of amacrine cells have long been studied, there is little understanding as to whether they have specific and distinct roles, rather than each serving a similar, general function. A substantial barrier to the understanding of inhibitory interneurons has been technical limitations on studying single cells among a diverse population. This proposal seeks to characterize the population of inhibitory amacrine cells using a novel approach to optically record the visual responses of the amacrine cell population while simultaneously recording populations of ganglion cells using an electrode array. This project focuses on responses to moving visual stimuli, which are ecologically important and critical to behavior and perception. The first aim of this proposal will measure the responses of a nearly complete amacrine population to moving stimuli and compare these with simultaneously recorded responses in the ganglion cell population. By measuring the similarity in responses between amacrine and ganglion cell populations, these experiments will test the hypothesis that amacrine cells are divided into two broad classes: one that resembles the more simple bipolar cell representation and one that is more tightly correlated with specific retinal ganglion cells. Te second aim of this proposal uses optical imaging and simultaneous intracellular and multielectrode recording to test the hypothesis that amacrine cells inhibit ganglion cells that the are correlated with under visual motion, despite the wide variation of preferred visual stimuli across amacrine cells. Finally, the third aim of this proposal will take advantage of our novel approach of directly perturbing individual interneurons intracellularly to test whether many types of amacrine cells act to reduce correlations in the ganglion cell population for different types of natural stimuli, thus creating an efficient representation of the visual scene. These studies will not only add to the knowledge of how an inhibitory population represents and transforms visual information, but will also test general principles applicable to all neural circuits. The results wll have immediate applicability to the emerging field of retinal prostheses. The objective of a retinal prosthesis system is to treat prevalent diseases such as age-related macular degeneration and retinitis pigmentosa by replacing the function of the damaged retina with a high-resolution electronic circuit. Measurements of the retinal neural code and the computations that are performed will be directly useful for incorporation into these prostheses systems.